Seoul National University NLP researcher: conversational search, LLM agents, and long-term dialogue memory
3
papers here
3
reviewed by Evolve
1
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-16) and cover the author's entire public record, not just the papers listed here. Reproducibility signal comes from Evolve's own reviews.
Research profile
Output—
The output axis is unscored: the snapshot provides no publication counts, venue breakdown, or productivity totals. This axis would measure the volume and consistency of research output; only three 2026 arXiv preprints are available, which is insufficient.
Impact—
The impact axis is unscored: no citation counts, h-index, or influence metrics appear in the data. It would measure how widely the community cites and builds on this work; all three papers are 2026 preprints, too recent for citation evidence.
Focus0/100
The focus score is 0 in the current snapshot. The underlying field distribution splits evenly across three labels — natural language processing (conversational AI), NLP/information retrieval, and NLP/IR with LLM agents — each at 0.333. Because these labels are thematically adjacent, the work appears topically tight (conversational search, agents, memory) even though the even three-way split drives the low computed score.
Collaboration58/100
The collaboration score is 58, based on 17 distinct coauthors in the available record. Twelve recur as frequent collaborators, including Kyomin Jung, Hyojeong Yu, Hyukhun Koh, Minsung Kim, Kang-il Lee, and Joongbo Shin, alongside international partners such as Abhisek Dash, Asmaa El Fraihi, Elisabeth Kirsten, and Qinyuan Wu. All three 2026 papers were team efforts, with Jang first author on one and a middle author on the other two.
Activity—
The activity axis is unscored: no per-year publication records or submission cadence data are available beyond the three 2026 catalog papers. It would measure research pace and engagement over time; the current single-year snapshot is insufficient to compute it.
Reproducibility43/100
The reproducibility score is 43, a mean across three scored papers. Evolve rates CompCQR as medium readiness (relies on LLM rank-1 selection and single-run evaluations), PRAGMA as medium (synthetic 400-query benchmark with judge circularity), and the cross-platform web search study as low (judge-model heterogeneity and confounds). The pattern suggests careful artifact-building with methodological caveats that still complicate independent replication.
Research summary
Yunah Jang is a natural language processing researcher at Seoul National University whose recent work sits at the intersection of conversational search, LLM agents, and long-term dialogue memory. The reviewed catalog contains three 2026 arXiv preprints. As first author, Jang led CompCQR, a training-free conversational query reformulation method that generates 325 permuted component queries, fuses them with reciprocal rank fusion, and applies LLM selection over dense retrieval and BM25; Evolve rates its reproducibility readiness as medium. Jang also co-authored PRAGMA, a carefully built 400-query benchmark showing that LLM memory systems fail at personalized guidance in lifelong conversations, and a cross-platform characterization of conversational LLM agents performing web search, described as the first measurement spanning four platforms. Bibliometric signals record 17 distinct coauthors (collaboration score 58) and a reproducibility mean of 43 across three scored papers, while the focus score of 0 reflects an even three-way split across closely related NLP/IR field labels. Output, impact, and activity metrics are not yet available. This is an early-stage profile: all evidence comes from 2026 preprints, with no citation or venue data yet recorded.
Collaboration pattern
Yunah Jang's collaboration score is 58, with 17 distinct coauthors in the available record. The recurring core includes Kyomin Jung, Hyojeong Yu, Hyukhun Koh, Minsung Kim, Kang-il Lee, and Joongbo Shin, alongside Seungeon Lee and international collaborators Abhisek Dash, Mahsa Amani, Asmaa El Fraihi, Elisabeth Kirsten, and Qinyuan Wu. All three reviewed 2026 papers were multi-author efforts; Jang led one as first author and contributed in middle positions on the others, with no corresponding-author role recorded.
Research keywords
Highlights
CompCQR: training-free conversational query reformulation (first author)
PRAGMA: 400-query benchmark exposing failures in LLM personalized guidance
First cross-platform measurement of agentic web search
Timeline
- 2026First-author CompCQR paper on training-free conversational query reformulation posted to arXiv.
- 2026Co-authored PRAGMA benchmark on personalized guidance and memory alignment in lifelong conversations.
- 2026Co-authored cross-platform characterization of web search by conversational LLM agents.
Papers in the Evolve catalog
Frequent co-authors
About this profile
This profile is AI-structured by Evolve.Pub from public bibliographic data (OpenAlex) and Evolve's own paper reviews. It is not written or endorsed by the researcher unless the profile is claimed. Radar dimensions are computed from real bibliometrics with fixed formulas — raw figures are shown alongside; missing dimensions mean insufficient data, never zero. Errors? Claim the profile to correct or hide it.
- Bibliometric scores are computed from a limited snapshot covering only the three reviewed 2026 arXiv preprints.
- Output, impact, and activity dimensions are unscored because no citation counts, publication totals, or yearly productivity data were provided.
- All catalog papers are arXiv preprints; no peer-reviewed venue information is present in the data.
- The profile is name-resolved from public bibliographic data (ORCID 0000-0002-2805-7530); the country field is missing in the source record.